Table 4. Multiple regression analysis of the impact of editorial leadership and of potential confounding parameters on journal quality, for 73 Italian and 69 UK journals that publish research on humans.

Predictors	SCImago journal rank		SCImago cites/doc		Impact factor
	Rb	R2c	Rb	R2c	Rb	R2c
Editorial leadershipa	0.609	37.1*	0.707	49.9†	0.707	49.9‡
+Publishing language	0.613	0.4	0.713	0.9	–d	–d
+International editorial board	0.616	0.8	0.738	4.5	0.721	2.0
+Country of origin	0.620	1.3	0.721	2.0	0.718	1.6
+Size (articles/year)	0.629	2.4	0.743	5.4	0.733	3.8
+% articles with international authorship	0.636	3.3	0.734	4.0	0.726	2.8
+All potential confounders	0.658	6.2	0.798	13.8	0.757	7.5
Total explained variance, %		43.3		63.7		57.4

Independent analyses were run for the criteria variables SCImago journal rank and SCImago cites/doc (available for 141 of the journals) and impact factor (75 journals).

^a All aspects but adherence to animal research laws were considered. ^b Multiple correlation coefficient for all predictors included in the model; coefficients for all models were statistically significant (p<0.0001). ^c Percentage of variance explained by inclusion of the new predictor in the model. ^d Language not entered into the model because the IF journal set has only one bilingual journal and no journals in Italian.

* F_(141,128) = 6.28; ^†F_(141,128) = 10.6; ^‡F_(75,62) = 5.15.